There's a huge reorg underway at M Science, the pioneer alt-data seller owned by Jefferies. It's shaking up its executive ranks and sales team, and cutting data scientist jobs.
- M Science has cut several people from its 20-person data science team this week, sources told Business Insider, and the firm's head of sales has been demoted to a part-time employee.
- The cuts to the data science team include former Palantir data scientist Dipanjan Sen, while former Citadel programmer Ajay Krishna is transitioning out of his role as head of data science and engineering.
- This year alone the firm has lost execs like its long-time head of sales Mark Trowbridge and head of strategy Daniel Entrup.
- Click here for more BI Prime stories.
A pioneer alternative-data seller is going through a major upheaval.
M Science's origins date back to 2002 as one of the first sellers of non-traditional data to financial firms in what has become a red-hot and fiercely crowded space.
Now, the company has cut several people from its 20-person data science team, several sources said, including Dipanjan Sen, who is the data science team lead and a former data scientist at Peter Thiel's Palantir. Ajay Krishna, head of data science and engineering at the firm and a former employee at Ken Griffin's Citadel, is transitioning out of his role.
Gerard Varjacques, the company's global head of sales, has transitioned to a part-time role, according to several sources, with M Science chief executive Mike Marrale picking up his responsibilities.
The firm has created a chief technology officer position for the first time, which will be filled by former Google employee Ben Tallman, who joined M Science in February as the head of engineering and infrastructure.
Varjacques was also hired in February after the departure of long-time head of sales Mark Trowbridge. The firm's head of strategy Daniel Entrup also left the firm this year.
Jefferies and M Science declined to comment.
Krishna led a pivot in the data science group when he joined M Science at the beginning of 2017, sources said, as the company began to build out a data science team separate from the research function. Krishna's team began to build its own data products to sell, most prominently SWIPE, a database that tracks consumer transactions and predicted spending habits using a machine-learning model.
M Science struggled to sell the product, a source familiar with the data set said, as competitors were already entrenched in the space. The source told Business Insider that SWIPE's data was also often delayed compared to competitors.
"We are changing the transaction data landscape with M Science's SWIPE," said Krishna in a statement in July 2018 announcing the launch. "Using extensive feature engineering and machine learning, we identify key drivers from the data to provide multiple viewpoints to our clients."
M Science has more than 15 years experience in the alt-data space
Established as Majestic Research in 2002, the company was one of the first to focus on the collection and analysis of non-traditional data sets, specifically credit card usage.
Cofounder and then-CEO of Majestic Research, Tony Berkman, is considered by some to be the godfather of alternative data. He now serves as a managing director at Two Sigma.
In October 2010, Majestic Research was acquired by brokerage ITG for $56 million. Less than six years later, ITG spun off Majestic Research, renamed M Science, to Leucadia National Corporation, a division of Jefferies, for $12 million, although it was not announced what parts of Majestic Research remained at ITG.
In recent years, M Science has doubled down on its efforts in the alternative-data space. The company had a string of product announcements in 2019, the most recent of which came in December and was a data set focused on the adoption of 5G and the devices connecting to the network.
But as investors, particularly quantitative hedge funds, have gotten more sophisticated in their digestion of alternative data sets, M Science's black-box type product has been a tougher sell, according to a second source familiar with the matter.
Quants found it difficult to incorporate the company's aggregated data sets into their models without knowing what was inside them to ensure there was no overlap with other data sets they were already using, the second source said.
The market for alternative data, which Deloitte projects will exceed $7 billion in 2020, has become crowded, and there have been some shakeouts already. Investors' increased demand for unique, non-traditional data sets has not necessarily guaranteed success for all of those in the space.
In September, Business Insider reported alternative-data company Thasos fired two-thirds of its staff and its then-CEO and cofounder, Greg Skibiski, stepped down after the firm faced issues making money selling to its main financial clients.
In addition to the competition that comes from new startups entering the space, traditional data companies are have also moving into alt data.
In 2018, Nasdaq acquired alternative-data company Quandl. Shortly thereafter, Bloomberg announced the launch of its own alternative-data store.
Got a tip? Contact Dan DeFrancesco via email at ddefrancesco@businessinsider.com, Signal (646-768-1650) or direct message on Twitter @dandefrancesco. Contact Bradley Saacks via email at bsaacks@businessinsider.com, Signal (919-816-5537) or direct message on Twitter @SaacksAttack.